In existing Radio Access Networks (RANs), such as Evolution Data—Optimized (EV-DO) networks, congestion is controlled using admission control, which involves decisions as to whether or not to allow new packet flows into the RAN while the RAN is congested. Disadvantageously, however, use of admission control to alleviate congestion is a slow process because only when new packet flows are established is there an opportunity to reduce congestion in the RAN.